[OE-core] [PATCH 0/6] Integrate Smart for RPM image generation

Mark Hatle mark.hatle at windriver.com
Sat Dec 8 17:10:41 UTC 2012


What is the build configuration for this chunk?

Also this looks like it might be something within the license class, not the 
rootfs, but I'm not sure.

--Mark

On 12/8/12 12:37 AM, Saul Wold wrote:
>
> There still seems to be a problem with the multilib build that I believe
> comes from this set of changes.
>
>> | Running intercept scripts:
>> | WARNING: The license listed MIT was not in the licenses collected for lib32-packagegroup-core-boot
>> | ls: cannot access /srv/home/pokybuild/yocto-autobuilder/yocto-slave/nightly-multilib/build/build/tmp/pkgdata/*/runtime-reverse/udev-extraconf: No such file or directory
>> | readlink: missing operand
>> | Try `readlink --help' for more information.
>> | basename: missing operand
>> | Try `basename --help' for more information.
>> | ERROR: Function failed: do_rootfs (see /srv/home/pokybuild/yocto-autobuilder/yocto-slave/nightly-multilib/build/build/tmp/work/qemux86_64-poky-linux/lib32-core-image-minimal/1.0-r0/temp/log.do_rootfs.14393 for further information)
>> NOTE: recipe lib32-core-image-minimal-1.0-r0: task do_rootfs: Failed
>> ERROR: Task 7 (virtual:multilib:lib32:/srv/home/pokybuild/yocto-autobuilder/yocto-slave/nightly-multilib/build/meta/recipes-core/images/core-image-minimal.bb, do_rootfs) failed with exit code '1'
>> NOTE: Tasks Summary: Attempted 1867 tasks of which 425 didn't need to be rerun and 1 failed.
>
> Sau!
>
>
> On 12/04/2012 11:49 AM, Mark Hatle wrote:
>> The following patch set relies on the previous "Misc fixes" sent out earlier
>> today.
>>
>> The set replaces the existing rootfs generation with one based on Smart.
>> The work was tested using buildhistory over a wide range of image types:
>> core-image-minimal
>> core-image-base
>> core-image-basic
>> core-image-lsb
>> core-image-sato
>> build-appliance
>>
>> Each of which contained minor changes from the former version, but the changes
>> can all be explained by a different package resolver.
>>
>> Multilib configurations of each of the above images were also verified on
>> qemux86_64 configuration with a few lib32 items added to the mix.  This was
>> shown, by buildhistory, to result in a more compact install solution.
>>
>> "populate_sdk" was also verified after this patch.  Buildhistory does not
>> verify SDKs, so these were verified by hand.  The target items all appeared
>> to match the expected image components.
>>
>> The incremental RPM installation code was verified as well.. This code was
>> verified in both multilib and non-multilib configurations.
>>
>> Finally, each of these were verified on: qemux86_64, qemumips, qemuppc, qemuarm,
>> and qemuarm configured w/ hardware float API.  This verifies that both the
>> ARM specific "gnueabi" issues and the _ to - in the arch names both work
>> properly.
>>
>> The following changes since commit 0c69e555dda2fa412ba12fa90886b0b92dc60e44:
>>
>>     license.bbclass: Avoid grep error message (2012-12-04 11:10:18 -0600)
>>
>> are available in the git repository at:
>>     git://git.yoctoproject.org/poky-contrib mhatle/smart
>>     http://git.yoctoproject.org/cgit.cgi//log/?h=mhatle/smart
>>
>> Mark Hatle (4):
>>     rootfs_rpm.bbclass: Revert to default shell logging behavior
>>     package_rpm: Update the way the multilib package names are translated
>>     package_rpm.bbclass: Add additional logging
>>     package_rpm.bbclass: Add support for incremental installs
>>
>> Paul Eggleton (2):
>>     classes/*_rpm: integrate Smart into RPM filesystem construction
>>     classes/rootfs_rpm: install smart instead of zypper in rpm-based
>>       images
>>
>>    meta/classes/package_rpm.bbclass      |  691 +++++++++++++--------------------
>>    meta/classes/populate_sdk_rpm.bbclass |   74 ++---
>>    meta/classes/rootfs_rpm.bbclass       |   86 ++---
>>    3 files changed, 338 insertions(+), 513 deletions(-)
>>





More information about the Openembedded-core mailing list